NMDS for Reference streams-3% removed
15 July, 2021
R version 4.0.3 (2020-10-10) – “Bunny-Wunnies Freak Out”
Packages used for NMDS: vegan (version 2.5-7)
Methods
The document serves as an example of analyses that will be conducted to identify natural variations in benthic communities across Virginia. These NMDS will support the Genus level IBI development process. This analysis is the first run of all of reference sites in Virginia. No West Virginia DEP data is used in this analysis. Several reference sites have been sent back to Virginia DEQ biologists for one more review.
The dataset used is a subset of reference stations collected in Virginia. If stations appeared in the dataset more than 4 times, then the most recent 4 samples were used and the rest removed. Taxa that occurred in the dataset < 3% of the time were removed. The data was log10 +1 transformed. Environmental factors were compiled for each station and used to plot over the NMDS to show environmental variation associated with the community matrix. The envfit function in Vegan was used to plot the continuous environmental variables.
The first step was to read in the reference site bug taxa list and environmental factors dataset for each station. Join the environmental dataset with the bug dataset to account for multiple observations of each station and collection date and time.
Check to make sure the bug and environmental join was successful:
Number of rows in Community Matrix: 854
Number or rows in Environmental Matrix: 858
The data was log10+1 transformed. Rare taxa (<=3%) were removed.
Run NMDS for reference communities
## Run 0 stress 0.1680336
## Run 1 stress 0.1684943
## ... Procrustes: rmse 0.00319579 max resid 0.06499605
## Run 2 stress 0.1682916
## ... Procrustes: rmse 0.002731399 max resid 0.06489706
## Run 3 stress 0.1680171
## ... New best solution
## ... Procrustes: rmse 0.002396556 max resid 0.068746
## Run 4 stress 0.1682642
## ... Procrustes: rmse 0.003469849 max resid 0.06873332
## Run 5 stress 0.1682643
## ... Procrustes: rmse 0.003465359 max resid 0.06877089
## Run 6 stress 0.1682404
## ... Procrustes: rmse 0.002568496 max resid 0.06477704
## Run 7 stress 0.1680358
## ... Procrustes: rmse 0.002385238 max resid 0.06871529
## Run 8 stress 0.1685203
## Run 9 stress 0.168241
## ... Procrustes: rmse 0.002540557 max resid 0.06483661
## Run 10 stress 0.1691233
## Run 11 stress 0.1680377
## ... Procrustes: rmse 0.002381471 max resid 0.06863401
## Run 12 stress 0.168083
## ... Procrustes: rmse 0.00283391 max resid 0.06877781
## Run 13 stress 0.1681116
## ... Procrustes: rmse 0.003007188 max resid 0.06880483
## Run 14 stress 0.1682477
## ... Procrustes: rmse 0.003289009 max resid 0.06874907
## Run 15 stress 0.1682993
## ... Procrustes: rmse 0.002937873 max resid 0.06479914
## Run 16 stress 0.1682408
## ... Procrustes: rmse 0.002518197 max resid 0.06480589
## Run 17 stress 0.168084
## ... Procrustes: rmse 0.002852725 max resid 0.0688647
## Run 18 stress 0.1682252
## ... Procrustes: rmse 0.002277517 max resid 0.06483624
## Run 19 stress 0.1682913
## ... Procrustes: rmse 0.003628881 max resid 0.06897773
## Run 20 stress 0.168242
## ... Procrustes: rmse 0.002564566 max resid 0.06479689
## Run 21 stress 0.1680831
## ... Procrustes: rmse 0.002833838 max resid 0.0687755
## Run 22 stress 0.1682715
## ... Procrustes: rmse 0.002732182 max resid 0.06484454
## Run 23 stress 0.1682402
## ... Procrustes: rmse 0.002551476 max resid 0.06478896
## Run 24 stress 0.169357
## Run 25 stress 0.1682819
## ... Procrustes: rmse 0.003475259 max resid 0.06897655
## Run 26 stress 0.1682244
## ... Procrustes: rmse 0.002268476 max resid 0.06483125
## Run 27 stress 0.1682643
## ... Procrustes: rmse 0.003465414 max resid 0.06869292
## Run 28 stress 0.1682656
## ... Procrustes: rmse 0.003467675 max resid 0.06870807
## Run 29 stress 0.168241
## ... Procrustes: rmse 0.002545915 max resid 0.06480415
## Run 30 stress 0.1680156
## ... New best solution
## ... Procrustes: rmse 0.0001815101 max resid 0.003862929
## ... Similar to previous best
## *** Solution reached##
## Call:
## metaMDS(comm = NMDSthree[, 6:144], k = 3, trymax = 1000)
##
## global Multidimensional Scaling using monoMDS
##
## Data: NMDSthree[, 6:144]
## Distance: bray
##
## Dimensions: 3
## Stress: 0.1680156
## Stress type 1, weak ties
## Two convergent solutions found after 30 tries
## Scaling: centring, PC rotation, halfchange scaling
## Species: expanded scores based on 'NMDSthree[, 6:144]'
Envfit results from Vegan package :
##
## ***VECTORS
##
## NMDS1 NMDS2 r2 Pr(>r)
## JulianDate -0.72223 0.69165 0.0743 0.660
## Latitude 0.90919 -0.41638 0.2239 0.258
## Longitude 0.89008 0.45580 0.2546 0.195
## US_L3CODE 0.23608 -0.97173 0.3200 0.128
## Order -0.87248 -0.48865 0.7801 0.001 ***
## EDASOrder -0.87248 -0.48865 0.7801 0.001 ***
## totalArea_sqMile -0.48524 -0.87438 0.4762 0.037 *
## ELEVMEAN -0.71649 -0.69759 0.5565 0.011 *
## SLPMEAN -0.24497 -0.96953 0.7939 0.001 ***
## wshdRain_mmyr -0.46527 -0.88517 0.4796 0.036 *
## siteRain_mmyr -0.59859 0.80106 0.2534 0.208
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Permutation: free
## Number of permutations: 999
##
## ***FACTORS:
##
## Centroids:
## NMDS1 NMDS2
## SeasonFall -0.6971 0.5872
## SeasonOutside Sample Window -0.5772 0.1178
## SeasonSpring -0.6572 0.2728
## GradientMACS -0.6686 0.3964
## CoastalTRUE -0.6686 0.3964
## US_L3NAMEMiddle Atlantic Coastal Plain -0.6506 0.6849
## US_L3NAMESoutheastern Plains -0.6785 0.2362
## US_L4NAMEChesapeake-Pamlico Lowlands and Tidal Marshes -0.6791 0.3657
## US_L4NAMEMid-Atlantic Flatwoods -0.6316 0.8978
## US_L4NAMERolling Coastal Plain -0.6785 0.2362
## ASSESS_REGPRO -0.6786 0.2597
## ASSESS_REGTRO -0.6316 0.8978
## TidalN -0.6668 0.4016
## TidalT -0.6791 0.3657
## VAHUSBCB -0.6071 0.5407
## VAHUSBCU -0.7208 0.5247
## VAHUSBYO -0.6517 0.0599
## BasinChes. Bay and Small Coastal Basin -0.6071 0.5407
## BasinChowan and Dismal Swamp River Basin -0.7208 0.5247
## BasinYork River Basin -0.6517 0.0599
## Basin_CodeChowan-Dismal -0.7208 0.5247
## Basin_CodeSmall Coastal -0.6071 0.5407
## Basin_CodeYork -0.6517 0.0599
## CountyCityNameCaroline -0.7296 -0.0188
## CountyCityNameDinwiddie -0.8387 0.3481
## CountyCityNameHanover -0.5737 0.1385
## CountyCityNameKing and Queen -0.5351 0.7157
## CountyCityNameNorthumberland -0.6791 0.3657
## CountyCityNameSouthampton -0.6316 0.8978
## CountyCityNameSussex -0.7526 -0.2416
## WQS_CLASSIII -0.7140 0.0904
## WQS_CLASSVII -0.6433 0.5664
## WQS_SPSTDS -0.6537 0.3536
## WQS_SPSTDSNEW-21 -0.8613 0.9534
## WQS_PWS -0.6686 0.3964
## WQS_TROUT -0.6686 0.3964
## WQS_TIER_III -0.6686 0.3964
##
## Goodness of fit:
## r2 Pr(>r)
## Season 0.1838 0.384
## Gradient 0.0000 1.000
## Coastal 0.0000 1.000
## US_L3NAME 0.2863 0.035 *
## US_L4NAME 0.4372 0.032 *
## ASSESS_REG 0.4251 0.006 **
## Tidal 0.0011 1.000
## VAHUSB 0.2941 0.126
## Basin 0.2941 0.126
## Basin_Code 0.2941 0.126
## CountyCityName 0.8722 0.002 **
## WQS_CLASS 0.3279 0.038 *
## WQS_SPSTDS 0.1648 0.267
## WQS_PWS 0.0000 1.000
## WQS_TROUT 0.0000 1.000
## WQS_TIER_III 0.0000 1.000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Permutation: free
## Number of permutations: 999
##
## 840 observations deleted due to missingnessPlot with Station IDs
Plot with Axis 3
Plot with Species
Season
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$Season, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## Fall Outside Sample Window Spring
## delta 0.673 0.6843 0.677
## n 414 13 427
##
## Chance corrected within-group agreement A: 0.03403
## Based on observed delta 0.6752 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Season with continuous variable vectors
Ecoregion
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$US_L3NAME, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## Blue Ridge Central Appalachians Middle Atlantic Coastal Plain
## delta 0.6029 0.612 0.5932
## n 154 40 25
## Northern Piedmont Piedmont Ridge and Valley Southeastern Plains
## delta 0.6522 0.6224 0.6621 0.6438
## n 101 138 259 137
##
## Chance corrected within-group agreement A: 0.08929
## Based on observed delta 0.6365 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Basin
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$Basin_Code, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## Appomattox Chowan-Dismal James-Lower James-Middle James-Upper New
## delta 0.5752 0.6644 0.6262 0.6185 0.6573 0.6587
## n 15 33 36 61 120 78
## Potomac-Lower Potomac-Shenandoah Rappahannock Roanoke Small Coastal
## delta 0.6925 0.6606 0.674 0.6185 0.5796
## n 31 42 144 87 21
## Tennessee-Big Sandy Tennessee-Clinch Tennessee-Holston Yadkin York
## delta 0.6108 0.6339 0.632 0.5781 0.6727
## n 16 50 53 4 63
##
## Chance corrected within-group agreement A: 0.07359
## Based on observed delta 0.6475 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Admin Region
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$ASSESS_REG, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## BRRO NRO PRO SWRO TRO VRO
## delta 0.6493 0.6978 0.6435 0.6405 0.5852 0.6483
## n 250 195 122 156 21 110
##
## Chance corrected within-group agreement A: 0.06113
## Based on observed delta 0.6562 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Sample Method
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$Gradient, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## MACS Riffle
## delta 0.6505 0.6619
## n 163 691
##
## Chance corrected within-group agreement A: 0.05608
## Based on observed delta 0.6598 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999BioRegion- Not in dataset
Stream Order
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$Order, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## 1 2 3 4 5 6
## delta 0.6754 0.6904 0.665 0.6638 0.6289 0.4322
## n 277 248 172 114 41 2
##
## Chance corrected within-group agreement A: 0.03666
## Based on observed delta 0.6733 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Water Quality Standard Class
##
## Call:
## mrpp(dat = bugsnms[, 6:144], grouping = samplescoresenv$WQS_CLASS, distance = "bray")
##
## Dissimilarity index: bray
## Weights for groups: n
##
## Class means and counts:
##
## III IV V VI VII
## delta 0.697 0.6508 0.6376 0.6115 0.6389
## n 331 185 87 209 42
##
## Chance corrected within-group agreement A: 0.05979
## Based on observed delta 0.6572 and expected delta 0.6989
##
## Significance of delta: 0.001
## Permutation: free
## Number of permutations: 999Natural Trout Water WQS?
